Autonomous Rover — Jetson ROS2 Platform

A personal project exploring real-time control and perception. Built on a Jetson Nano using ROS2 with onboard cameras for obstacle detection.

ROS2Jetson PIDOpenCVPyTorch
Autonomous Rover

Overview

Built a four-wheel rover integrating sensor data and control loops to track paths autonomously. Implemented image-based obstacle avoidance and data visualization for debugging.

Key Work

  • Implemented PID control and trajectory planning ( 25 % faster loop response ).
  • Trained lightweight neural nets in PyTorch for obstacle classification.
  • Automated simulation tests and tuned gains through iterative field runs.
  • Developed live telemetry plots for stability and error analysis.